Neumarker and Cortica partner to develop AI biomarkers for autism
Neumarker, Inc. and Cortica, Inc. have announced a strategic collaboration to apply AI-driven EEG analysis to autism care, combining Neumarker's large-scale brain signal modelling with Cortica's de-identified longitudinal clinical dataset covering more than 20,000 paediatric patients annually.
The partnership centres on a shared ambition to identify neurological biomarkers capable of predicting how individual children with autism spectrum disorder respond to different therapeutic interventions. Clinicians currently lack objective, data-driven tools for stratifying patients before treatment, meaning intervention choices are largely empirical. Both companies describe this variability in treatment response as the core scientific problem the collaboration aims to address.
What each party brings
Neumarker is a venture-backed company based in Mountain View, California, whose core asset is an AI-native platform built around a large-scale electroencephalography foundation model. Its lead programme, NMK-001, is in development for epilepsy, a separate CNS indication that has nonetheless generated the training data and model architecture the company now intends to extend into autism. Cortica, headquartered in San Diego, operates as a physician-led, whole-child care provider and employs more than 2,300 clinicians across its network. It holds what it describes as one of the largest longitudinal datasets in neurodevelopmental care, accumulated through its integrated model combining medical, applied behaviour analysis, and developmental services.
Ken Wang, co-founder and chief executive of Neumarker, framed the collaboration as an effort to provide clinicians with objective tools that are currently absent from the field. "Children with autism can respond differently to the same interventions, yet clinicians today have limited objective tools to help predict treatment response," he said. "We hope to identify biomarkers that can support a more personalised approach to care."
The collaboration will operate on de-identified data and, according to the companies, in compliance with applicable patient privacy and research requirements. No financial terms were disclosed, and no specific research timeline or publication commitment was announced.
Market context and competitive landscape
Interest in AI-assisted biomarker discovery for neurodevelopmental conditions has grown substantially over the past several years, driven by the recognition that autism represents a highly heterogeneous clinical population poorly served by one-size-fits-all protocols. EEG has attracted particular attention as a relatively low-cost, scalable modality that can generate high-dimensional brain activity data without the access constraints of MRI or PET imaging. A number of academic groups and early-stage companies are pursuing computational approaches to EEG-based stratification in autism, though none has yet reached regulatory approval for a companion diagnostic in the indication.
The partnership's commercial logic is clear: Neumarker gains access to a rare, large-scale real-world dataset of neurodevelopmental patients, while Cortica can position its innovation network as a site for translational research that may eventually inform its own clinical protocols. Whether the collaboration produces a validated, clinically deployable biomarker will depend on the size and quality of the EEG data that can be gathered within Cortica's care settings, and on regulatory clarity around what constitutes an approvable biomarker endpoint in autism.
Regulatory read-across from the FDA's work on biomarker qualification in other CNS areas, including its guidance on Alzheimer's disease biomarkers, is likely to shape how the companies design any future submission strategy. The agency has historically required prospective validation data for biomarker qualification, a bar that retrospective longitudinal datasets alone may not clear.
Investors and clinical observers will be watching for the publication of research outputs from the Cortica Innovation Network, and for any signal that NMK-001's epilepsy programme is generating the kind of model performance that credibly transfers to autism.